XMR-stak-JK 2.10.7 Compiled with no devfee (Updated: 2019-8-18)

[neemzsoul]

Mining loki here - How to active Double Threads for RX 580 8gb? Thanks for the miner

[1715061013]

1715061013

[1715061013]

1715061013

[zenixchen]

Can you help to compile new XMR stak 2.4.5 0% dev for new fork from Masai and Haven ? Thanks a lot

[JaredKaragen]

Can you help to compile new XMR stak 2.4.5 0% dev for new fork from Masai and Haven ? Thanks a lot

https://imgur.com/a/EO5Ru0D

CUDA is being a @#$!#@$ again.    Its saying its not a compatible version of the compiler;  yet it is.  I have updated nothing since the last 2.4.4 build.

IDK what gives.  This is the same sort of crap that happened to me last time.

Mining loki here - How to active Double Threads for RX 580 8gb? Thanks for the miner

you will have to go on the xmr-stak github and look through the documentation.   I personally don't understand the json config parameters as to how to create them as multiple threads.... so I would refer you to the original docs.

[MagicSmoker]

Mining loki here - How to active Double Threads for RX 580 8gb? Thanks for the miner

You simply duplicate the lines starting with { "index" in the amd.txt file like this:

Code:

"gpu_threads_conf" : [
  // gpu: Ellesmere memory:3712
  // compute units: 32
  { "index" : 0, "intensity" : 848, "worksize" : 8, "affine_to_cpu" : false, "strided_index" : 2, "mem_chunk" : 2, "comp_mode" : true },
  { "index" : 0, "intensity" : 848, "worksize" : 8, "affine_to_cpu" : false, "strided_index" : 2, "mem_chunk" : 2, "comp_mode" : true },
],

This is for a single RX 570 (4GB) because that's what I have.

The easiest way to do this is to let xmr-stak generate the amd.txt file automatically (which it will do the first time it is run) then edit that file to double up on the index lines. You will likely have to adjust the intensity value to get the best hashrate and stop it from crashing, but the other parameters should be the same as shown above (the default choices for strided_index and mem_chunk may not be ideal).

NB - I'm still running 2.4.2.

[btckesh]

maybe worth a look when i update my hardware later this year. thanks

[UpStakeLover]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

[JaredKaragen]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

[newmz]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

I'm not sure what the hell happens (because so far I have been asleep each time) but I have some issue which seems to crash three instances of the miner (2.4.4) running on two machines. Last night it was after running great for 16 hours, mining Loki (which has a very low difficulty right now which is good, but has taken a price hit - but I am HODLing anyway).

Did you manage to compile 2.4.5 yet? I was going to try to teach myself how sometime but it's like going back to f'ing school.....

[newmz]

Eventually I solved this issue by running separate instances of XMR-stak. 1st one running the 1080 and the 1st 2 1070s plus cpu core 0, second one running the remaining two 1070s and cpu core 2. With this setup, now finally all five cards are at 650-700h/s. It took quite a bit of messing with the configs of the second instance to get it to even run, and I have to set it to verbosity 4 and 30secs because pressing h doesn't respond and give me hash-rate.

If anyone else is running a 4 or more GPU rig (especially if you use 1070s/1080s) and you aren't getting full hash rate on all GPUs I suggest trying what I did - it works great.
Like I said, getting the second instance of xmr-stak running on the same machine means changing quite a few things in the configuration.txt so if anyone is interested I will try to remember what I did.

Great tip, it should be on a top 10 list. I 'accidentally' found this too with my rig with mis-matched cards, that running different instances of the miner for each card(s) helps a lot.   

so are you running an individual instance for every card? does that give you more stability? I have divided the work up into 1 CPU core & 3 GPUs on one instance and 1 CPU core 2 and 2 GPUs on a second instance, but like I said in my previous post - it all still seems to crash eventually. Sometimes it manages to live for a long time and other times only a few hours, but this is also happening on a second machine with only 3 GPUs (one miner instance). I was beginning to suspect that it's OhGodAnEthLargementPill I am running for better 1080 performance rather than xmr-stak that is the issue, but I don't know. I guess I need to make it keep a log and dig into that.

It seems that the majority of cryptonight miners are more AMD focussed so I guess there is a lack of stability, but unfortunately the only ccminer that does cryptonight hasn't updated to the newer ASIC restistant variations of the algorithm like cryptonight-heavy.

[JaredKaragen]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

I'm not sure what the hell happens (because so far I have been asleep each time) but I have some issue which seems to crash three instances of the miner (2.4.4) running on two machines. Last night it was after running great for 16 hours, mining Loki (which has a very low difficulty right now which is good, but has taken a price hit - but I am HODLing anyway).

Did you manage to compile 2.4.5 yet? I was going to try to teach myself how sometime but it's like going back to f'ing school.....

no luck on trying with 2.4.5 yet.  Probably sometime tomorrow since Ill be at a location where I can uninstall and reinstall everything for the third time to try and compile it.

The issue was a fluke;  seems right before the completion of the new windows 18xx update.   I tried running 2.4.2, and it gave the same no connect error.

When I came home later that day my machine had rebooted itself automatically, and 2.4.4 launched no problems again.   It was an odd one.


Go into the CPU config, and remove the CPU threads.... leave it blank brackets {}.   See if stability comes back.   I have a LGA775 machine right now that gets TLS socket errors and reconnects every so often.  The auto emails from nanopool from that machine are quite annoying.

[onionpealer]

Eventually I solved this issue by running separate instances of XMR-stak. 1st one running the 1080 and the 1st 2 1070s plus cpu core 0, second one running the remaining two 1070s and cpu core 2. With this setup, now finally all five cards are at 650-700h/s. It took quite a bit of messing with the configs of the second instance to get it to even run, and I have to set it to verbosity 4 and 30secs because pressing h doesn't respond and give me hash-rate.

If anyone else is running a 4 or more GPU rig (especially if you use 1070s/1080s) and you aren't getting full hash rate on all GPUs I suggest trying what I did - it works great.
Like I said, getting the second instance of xmr-stak running on the same machine means changing quite a few things in the configuration.txt so if anyone is interested I will try to remember what I did.

Great tip, it should be on a top 10 list. I 'accidentally' found this too with my rig with mis-matched cards, that running different instances of the miner for each card(s) helps a lot.   

so are you running an individual instance for every card? does that give you more stability? I have divided the work up into 1 CPU core & 3 GPUs on one instance and 1 CPU core 2 and 2 GPUs on a second instance, but like I said in my previous post - it all still seems to crash eventually. Sometimes it manages to live for a long time and other times only a few hours, but this is also happening on a second machine with only 3 GPUs (one miner instance). I was beginning to suspect that it's OhGodAnEthLargementPill I am running for better 1080 performance rather than xmr-stak that is the issue, but I don't know. I guess I need to make it keep a log and dig into that.

It seems that the majority of cryptonight miners are more AMD focussed so I guess there is a lack of stability, but unfortunately the only ccminer that does cryptonight hasn't updated to the newer ASIC restistant variations of the algorithm like cryptonight-heavy.

Well, no, not for every card, only if one card looked like it needed special treatment.
I also ran different instances of a miner if I wanted to spread the risk and mine different coins on the same rig.

[redscream]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

I'm not sure what the hell happens (because so far I have been asleep each time) but I have some issue which seems to crash three instances of the miner (2.4.4) running on two machines. Last night it was after running great for 16 hours, mining Loki (which has a very low difficulty right now which is good, but has taken a price hit - but I am HODLing anyway).

Did you manage to compile 2.4.5 yet? I was going to try to teach myself how sometime but it's like going back to f'ing school.....

no luck on trying with 2.4.5 yet.  Probably sometime tomorrow since Ill be at a location where I can uninstall and reinstall everything for the third time to try and compile it.

The issue was a fluke;  seems right before the completion of the new windows 18xx update.   I tried running 2.4.2, and it gave the same no connect error.

When I came home later that day my machine had rebooted itself automatically, and 2.4.4 launched no problems again.   It was an odd one.


Go into the CPU config, and remove the CPU threads.... leave it blank brackets {}.   See if stability comes back.   I have a LGA775 machine right now that gets TLS socket errors and reconnects every so often.  The auto emails from nanopool from that machine are quite annoying.

I really appreciate if you can compile the latest version 2.4.5. Thanks

[MagicSmoker]

I really appreciate if you can compile the latest version 2.4.5. Thanks

Read post #92, ya lazy newb.

[JaredKaragen]

I really appreciate if you can compile the latest version 2.4.5. Thanks

Read post #92, ya lazy newb.

lol, yeah, ive been under the weather the past week and my life has definitely not been normal.  I am hoping to get into work today and work late to try and make up for lost time.

As soon as I am able to sit down at a high speed internet location, ill attempt another recompile.


I noticed with 2.4.4 my cuda libraries would not be able to use the old gtx 5xx series and older.... no bueno.   I may have to remove cuda 9, and reinstall cuda 7.5 and cuda 8, and see if I can get support back in the 4.4.5 version.    I attempted swapping for my old cuda_backend.dll file that worked; and it crashed the app, so I am assuming that happens because the new algos and functions added to cuda_backend.dll were moved around and changed somehow making the old library foreign language to the new app build.


Ill be walking into uncharted territory, because I believe to compile with cuda 7.5 and 8, Ill need to install VS2015....   so yeah.   fingers crossed I guess.

I just wish this crap would continue working after the first time you reboot the machine.

Not helping that im doing it on a laptop, but I ritually hibernate it now.  just cause.   

[redscream]

I really appreciate if you can compile the latest version 2.4.5. Thanks

Read post #92, ya lazy newb.

lol, yeah, ive been under the weather the past week and my life has definitely not been normal.  I am hoping to get into work today and work late to try and make up for lost time.

As soon as I am able to sit down at a high speed internet location, ill attempt another recompile.


I noticed with 2.4.4 my cuda libraries would not be able to use the old gtx 5xx series and older.... no bueno.   I may have to remove cuda 9, and reinstall cuda 7.5 and cuda 8, and see if I can get support back in the 4.4.5 version.    I attempted swapping for my old cuda_backend.dll file that worked; and it crashed the app, so I am assuming that happens because the new algos and functions added to cuda_backend.dll were moved around and changed somehow making the old library foreign language to the new app build.


Ill be walking into uncharted territory, because I believe to compile with cuda 7.5 and 8, Ill need to install VS2015....   so yeah.   fingers crossed I guess.

I just wish this crap would continue working after the first time you reboot the machine.

Not helping that im doing it on a laptop, but I ritually hibernate it now.  just cause.   

Crossed my fingers... I wish I would help you...  hope you will get it fixed...

[stevascha]

nice thread here, looking forward 2.4.5 version

[redscream]

nice thread here, looking forward 2.4.5 version

I have complied for amd and running with my vegas on my farm. I will share with you folks tomorrow when I make sure the stability.

[FFI2013]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

I'm not sure what the hell happens (because so far I have been asleep each time) but I have some issue which seems to crash three instances of the miner (2.4.4) running on two machines. Last night it was after running great for 16 hours, mining Loki (which has a very low difficulty right now which is good, but has taken a price hit - but I am HODLing anyway).

Did you manage to compile 2.4.5 yet? I was going to try to teach myself how sometime but it's like going back to f'ing school.....

no luck on trying with 2.4.5 yet.  Probably sometime tomorrow since Ill be at a location where I can uninstall and reinstall everything for the third time to try and compile it.

The issue was a fluke;  seems right before the completion of the new windows 18xx update.   I tried running 2.4.2, and it gave the same no connect error.

When I came home later that day my machine had rebooted itself automatically, and 2.4.4 launched no problems again.   It was an odd one.


Go into the CPU config, and remove the CPU threads.... leave it blank brackets {}.   See if stability comes back.   I have a LGA775 machine right now that gets TLS socket errors and reconnects every so often.  The auto emails from nanopool from that machine are quite annoying.

I got sick of windows sneaking in updates I had all my rigs running 10 pro with updates totally disabled and somehow they all updated so after playing on google for awhile I was able to find a win 10 pro stripped down with no apps or anything I haven't had one issue since

[newmz]

thanks for making this. I tried to follow instructions, but I had trouble with visual studio since I already have a few versions of it installed, the cuda downloader for visual studio integration failed.

Yours worked instantly! 2 ez 

you are welcome.


I have expeirenced a very strange connect error to nanopool on only one of my machines....  Anyone else getting this issue?

I'm not sure what the hell happens (because so far I have been asleep each time) but I have some issue which seems to crash three instances of the miner (2.4.4) running on two machines. Last night it was after running great for 16 hours, mining Loki (which has a very low difficulty right now which is good, but has taken a price hit - but I am HODLing anyway).

Did you manage to compile 2.4.5 yet? I was going to try to teach myself how sometime but it's like going back to f'ing school.....

no luck on trying with 2.4.5 yet.  Probably sometime tomorrow since Ill be at a location where I can uninstall and reinstall everything for the third time to try and compile it.

The issue was a fluke;  seems right before the completion of the new windows 18xx update.   I tried running 2.4.2, and it gave the same no connect error.

When I came home later that day my machine had rebooted itself automatically, and 2.4.4 launched no problems again.   It was an odd one.


Go into the CPU config, and remove the CPU threads.... leave it blank brackets {}.   See if stability comes back.   I have a LGA775 machine right now that gets TLS socket errors and reconnects every so often.  The auto emails from nanopool from that machine are quite annoying.

I got sick of windows sneaking in updates I had all my rigs running 10 pro with updates totally disabled and somehow they all updated so after playing on google for awhile I was able to find a win 10 pro stripped down with no apps or anything I haven't had one issue since

I was also sick of win 10 updates and all that "telemetry" running. I now use win 8.1 on all my rigs. xmr-stak runs OK for Nvidia GPU mining - I get up to 740h/s per gtx1070 (30 threads, 60 blocks, core 0, memory +750, powerlimit 75%).

@JaredKaragen - did you ever manage to compile 2.4.5? I really need to teach myself all that is necessary to compile for Windows or even maybe run linux.

[tbearhere]

@JaredKaragen   Did you ever compile XMR-STAK-2.4.8 ?  Thx